We applaud Heintz & Scott-Phillips's (H&S-P's) argument that the gulf between human communication and that of other animals arises primarily from the astonishing power of human social and pragmatic reasoning. We agree, too, that the unique flexibility and sophistication of natural language, in contrast to nonhuman animal communication systems, arise from a suite of cognitive abilities underlying such reasoning, rather than from any human-specific “universal grammar,” encoding abstract syntactic knowledge.
From a pragmatics-first perspective, however, the question remains: What is the route from non-linguistic communication, driven by a powerful “pragmatic engine,” to the creation of the astonishing complexity of full-blown combinatorial language? In this commentary, we argue that the game of charades provides a window not only into the nature of human pragmatic inference, but also into how linguistic systems can begin to emerge through a process of conventionalization (Christiansen & Chater, Reference Christiansen and Chater2022). We suggest, moreover, that processes of cultural evolution, without further biological evolution, can lead to the creation of a full-blown language, with the spontaneous, although partial, emergence of complex syntax.
To fix our intuitions, consider a charade aimed at conveying The Hound of the Baskervilles, first by miming the act of peering through a magnifying glass (hoping to bring to mind Sherlock Holmes) and then imitating a dog-like baying and biting action (to bring to mind the hound). While H&S-P focus on the complementarity between mechanisms for expression and interpretation of communicative signals, we stress that successful integration of such mechanism also requires communication to be a collaborative process (see Brennan & Clark, Reference Brennan and Clark1996; Clark, Reference Clark1996; Misyak & Chater, Reference Misyak and Chater2022). Thus, miming looking through a magnifying glass will only be taken to convey Holmes if the existence of the relevant association is common to all participants. Similarly, the relevance of Holmes to the target book title requires knowing that The Hound of the Baskervilles is a Sherlock Holmes mystery. If the observer doesn't know this then the communicative signal will likely fail. More generally, successful improvised communication requires all parties implicitly agreeing, given their common knowledge and goals, on a particular mapping between signals and meanings. Whatever the actor intends the charade to convey, the charade only succeeds in doing so if everyone involved interprets the charade in the same way (or closely enough for their communicative goal to be achieved). The capacity for establishing common ground, and engaging in joint reasoning in light of that common ground, is arguably crucial for coordinated social behavior of all kinds, and it is particularly central to the coordination of signal-meaning mappings underlying communication.
Charades are, of course, typically one-offs; and the charm of the game is the continual need for ingenuity and creativity from all players. But if the game is played repeatedly by the same people, conventions can rapidly become established. Thus, the magnifying glass gesture may become increasingly simplified and stylized, and its use broadened to convey detectives of all kinds, crime stories and movies, actual crimes, and so on. More generally, each new charade can build, in arbitrarily creative ways, upon the common ground of prior charades.
We have recently argued (Christiansen & Chater, Reference Christiansen and Chater2022) that the gradual conventionalization of charades captures, in miniature, some crucial aspects of the cultural evolution of language. The linguistic signal becomes increasingly standardized and simplified over time; and the meanings conveyed can both sprawl in many directions. Thus, everyday words, such as game, set, or shallow have endless interlocking meanings but, as Wittgenstein (Reference Wittgenstein1953) stressed, with no common definitional core (e.g., consider shallow waters, slopes, boats, bowls, spoons, thoughts, etc.).
The process of erosion and simplification of form, and broadening of meaning, parallels the process of grammaticalization widely observed in comparative and historical linguistics (e.g., Bybee, Perkins & Pagliuca, Reference Bybee, Perkins and Pagliuca1994; Hopper & Traugott, Reference Hopper and Traugott2003). Grammaticalization is the process by which some “content” words become so stereotyped in use, and so “bleached” of meaning, that they take on purely grammatical functions. Thus, for example, the content verb to will has in English also taken on a purely grammatical function (e.g., I will eat shifts from signaling an intention to eat, which must necessarily happen in the future, to a pure future-tense marker, irrespective of intention, as in the temperature will rise). Processes of simplification and erosion can also cause distinct words to collapse together, to create morphological complexity (thus, forms of to have have joined with verb stems to mark the future tense in many Romance languages) (Coleman, Reference Coleman1971; Fleischman, Reference Fleischman1982). The creation of grammatical words and functions and the increasing standardization of their use provides the starting point for complex syntactic patterning.
The linguistic signal consists of recycled parts with partially conventionalized meanings, although always with the possibility of new and often highly creative uses (Contreras Kallens & Christiansen, Reference Contreras Kallens and Christiansen2022). Thus, we continually extend meanings using rich pragmatic inference, such as in metonymy (e.g., take this drink to the pancakes by the window – where the pancakes substitutes for the customer with the pancakes) and extend meanings across domains by elaborate and partially consistent processes of metaphor (e.g., famously mappings between physical and mental objects and transportation [Lakoff & Johnson, Reference Lakoff and Johnson1980], so that we can give a person an idea, leave a worry behind, have it at the back of one's mind, etc.). Thus, the creative charade-like process remains at the heart of linguistic communication, but built on a system of conventions that has become entrenched over generations of language use.
The process of grammaticalization is, we suggest, part of the broader process of cultural evolution of language – by which linguistic forms and their meanings are continually reshaped by the multiple constraints of our perceptual, motor, and cognitive machinery, as well as the continually changing communicative challenges that we face (Christiansen & Chater, Reference Christiansen and Chater2022). Moreover, different linguistic conventions will continually be shaping each other, through processes of similarity, analogy, and competition. If H&S-P are right, and cognitive pragmatics is prior to, and underpins, linguistic communication, it is natural to consider the patterns exhibited by natural languages not as arising from a distinctive special-purpose biological endowment for syntax (Berwick & Chomsky, Reference Berwick and Chomsky2016), but through a process of spontaneous order over generations of cultural evolution (Chater & Christiansen, Reference Chater, Christiansen, Lappin and Bernady2022).
We applaud Heintz & Scott-Phillips's (H&S-P's) argument that the gulf between human communication and that of other animals arises primarily from the astonishing power of human social and pragmatic reasoning. We agree, too, that the unique flexibility and sophistication of natural language, in contrast to nonhuman animal communication systems, arise from a suite of cognitive abilities underlying such reasoning, rather than from any human-specific “universal grammar,” encoding abstract syntactic knowledge.
From a pragmatics-first perspective, however, the question remains: What is the route from non-linguistic communication, driven by a powerful “pragmatic engine,” to the creation of the astonishing complexity of full-blown combinatorial language? In this commentary, we argue that the game of charades provides a window not only into the nature of human pragmatic inference, but also into how linguistic systems can begin to emerge through a process of conventionalization (Christiansen & Chater, Reference Christiansen and Chater2022). We suggest, moreover, that processes of cultural evolution, without further biological evolution, can lead to the creation of a full-blown language, with the spontaneous, although partial, emergence of complex syntax.
To fix our intuitions, consider a charade aimed at conveying The Hound of the Baskervilles, first by miming the act of peering through a magnifying glass (hoping to bring to mind Sherlock Holmes) and then imitating a dog-like baying and biting action (to bring to mind the hound). While H&S-P focus on the complementarity between mechanisms for expression and interpretation of communicative signals, we stress that successful integration of such mechanism also requires communication to be a collaborative process (see Brennan & Clark, Reference Brennan and Clark1996; Clark, Reference Clark1996; Misyak & Chater, Reference Misyak and Chater2022). Thus, miming looking through a magnifying glass will only be taken to convey Holmes if the existence of the relevant association is common to all participants. Similarly, the relevance of Holmes to the target book title requires knowing that The Hound of the Baskervilles is a Sherlock Holmes mystery. If the observer doesn't know this then the communicative signal will likely fail. More generally, successful improvised communication requires all parties implicitly agreeing, given their common knowledge and goals, on a particular mapping between signals and meanings. Whatever the actor intends the charade to convey, the charade only succeeds in doing so if everyone involved interprets the charade in the same way (or closely enough for their communicative goal to be achieved). The capacity for establishing common ground, and engaging in joint reasoning in light of that common ground, is arguably crucial for coordinated social behavior of all kinds, and it is particularly central to the coordination of signal-meaning mappings underlying communication.
Charades are, of course, typically one-offs; and the charm of the game is the continual need for ingenuity and creativity from all players. But if the game is played repeatedly by the same people, conventions can rapidly become established. Thus, the magnifying glass gesture may become increasingly simplified and stylized, and its use broadened to convey detectives of all kinds, crime stories and movies, actual crimes, and so on. More generally, each new charade can build, in arbitrarily creative ways, upon the common ground of prior charades.
We have recently argued (Christiansen & Chater, Reference Christiansen and Chater2022) that the gradual conventionalization of charades captures, in miniature, some crucial aspects of the cultural evolution of language. The linguistic signal becomes increasingly standardized and simplified over time; and the meanings conveyed can both sprawl in many directions. Thus, everyday words, such as game, set, or shallow have endless interlocking meanings but, as Wittgenstein (Reference Wittgenstein1953) stressed, with no common definitional core (e.g., consider shallow waters, slopes, boats, bowls, spoons, thoughts, etc.).
The process of erosion and simplification of form, and broadening of meaning, parallels the process of grammaticalization widely observed in comparative and historical linguistics (e.g., Bybee, Perkins & Pagliuca, Reference Bybee, Perkins and Pagliuca1994; Hopper & Traugott, Reference Hopper and Traugott2003). Grammaticalization is the process by which some “content” words become so stereotyped in use, and so “bleached” of meaning, that they take on purely grammatical functions. Thus, for example, the content verb to will has in English also taken on a purely grammatical function (e.g., I will eat shifts from signaling an intention to eat, which must necessarily happen in the future, to a pure future-tense marker, irrespective of intention, as in the temperature will rise). Processes of simplification and erosion can also cause distinct words to collapse together, to create morphological complexity (thus, forms of to have have joined with verb stems to mark the future tense in many Romance languages) (Coleman, Reference Coleman1971; Fleischman, Reference Fleischman1982). The creation of grammatical words and functions and the increasing standardization of their use provides the starting point for complex syntactic patterning.
The linguistic signal consists of recycled parts with partially conventionalized meanings, although always with the possibility of new and often highly creative uses (Contreras Kallens & Christiansen, Reference Contreras Kallens and Christiansen2022). Thus, we continually extend meanings using rich pragmatic inference, such as in metonymy (e.g., take this drink to the pancakes by the window – where the pancakes substitutes for the customer with the pancakes) and extend meanings across domains by elaborate and partially consistent processes of metaphor (e.g., famously mappings between physical and mental objects and transportation [Lakoff & Johnson, Reference Lakoff and Johnson1980], so that we can give a person an idea, leave a worry behind, have it at the back of one's mind, etc.). Thus, the creative charade-like process remains at the heart of linguistic communication, but built on a system of conventions that has become entrenched over generations of language use.
The process of grammaticalization is, we suggest, part of the broader process of cultural evolution of language – by which linguistic forms and their meanings are continually reshaped by the multiple constraints of our perceptual, motor, and cognitive machinery, as well as the continually changing communicative challenges that we face (Christiansen & Chater, Reference Christiansen and Chater2022). Moreover, different linguistic conventions will continually be shaping each other, through processes of similarity, analogy, and competition. If H&S-P are right, and cognitive pragmatics is prior to, and underpins, linguistic communication, it is natural to consider the patterns exhibited by natural languages not as arising from a distinctive special-purpose biological endowment for syntax (Berwick & Chomsky, Reference Berwick and Chomsky2016), but through a process of spontaneous order over generations of cultural evolution (Chater & Christiansen, Reference Chater, Christiansen, Lappin and Bernady2022).
Financial support
NC was supported by the ESRC Network for Integrated Behavioural Science (grant number ES/K002201/1).
Conflict of interest
None.